Lung cancer radiation therapy in Australia and New ...

Journal of Medical Imaging and Radiation Oncology (2016) –

Lung cancer radiation therapy in Australia and New Zealand: Patterns of practice Syed Muntasser Islam,1 Shalini K Vinod,2 Margot Lehman,3 Shankar Siva,4 Tomas Kron,5 Patrick M Dwyer,6 Lois Holloway,7,8,9 Louis Lao,10,11,12 Mei Ling Yap9,13,14 and Jeremy D Ruben1,15 1 Radiation Oncology, William Buckland Radiotherapy Centre, Melbourne, Victoria, Australia 2 Cancer Therapy Centre, Liverpool Hospital, Liverpool BC, New South Wales, Australia 3 Radiation Oncology, Princess Alexandra Hospital, Brisbane, Queensland, Australia 4 Radiation Oncology, Peter MacCallum Cancer Centre, East Melbourne, Victoria, Australia 5 Medical Physics, Peter MacCallum Cancer Centre, East Melbourne, Victoria, Australia 6 North Coast Cancer Institute, Lismore, New South Wales, Australia 7 Medical Physics, Liverpool Hospital, Liverpool BC, New South Wales, Australia 8 South Western Clinical School, University of New South Wales, Sydney, New South Wales, Australia 9 Ingham Institute for Applied Medical Research, Liverpool BC, New South Wales, Australia 10 Radiation Oncology, Auckland City Hospital, Auckland, New Zealand 11 Auckland Radiation Oncology, Auckland, New Zealand 12 University of Auckland, Auckland, New Zealand 13 Radiation Oncology, Liverpool Hospital, Liverpool BC, New South Wales, Australia 14 Radiation Oncology, Macarthur Cancer Therapy Centre, Western Sydney University, Campbelltown, New South Wales, Australia 15 Monash University, Melbourne, Victoria, Australia

SM Islam MBBS; SK Vinod MBBS, MD, FRANZCR; M Lehman MBBS, FRANZCR, GDPH; S Siva PhD, MBBS, FRANZCR; T Kron PhD; PM Dwyer MBBS, FRANZCR; L Holloway PhD; L Lao MBChB; ML Yap MBBS, BSc, FRANZCR; JD Ruben MD, MBBCh(Hons), FCRadOnc, FRANZCR, Mmed. Correspondence Dr Syed M Islam, William Buckland Radiotherapy Centre, The Alfred, Commercial Rd, Prahran, Vic. 3181, Australia. Email: [email protected] Conflict of interest: None. Submitted 21 December 2015; accepted 24 April 2016. doi:10.1111/1754-9485.12475

Abstract Introduction: The RANZCR Faculty of Radiation Oncology Lung Interest Cooperative (FROLIC) surveyed patterns of lung cancer radiation therapy practice for non small cell (NSCLC) and small cell lung cancer (SCLC) to evaluate current patterns of care and potential for improvement. Methods: In October 2014, Radiation Oncologists (ROs) from all 62 departments in Australia and New Zealand were invited to a web-based survey directed at those treating lung cancer. Questions covered current radiation therapy practice as well as quality measures. Results: Fifty-eight per cent of respondents used 4D-CT simulation. For curative treatment, 98% employed 3D-CRT and 34% intensity modulated radiotherapy (IMRT) techniques. Treatment verification was primarily performed using cone-beam CT (86%). In NSCLC, the commonest curative dose-fractionation regime was 60 Gy/30# (96%) and for palliative intent, 30 Gy/10# (76%). Forty-four per cent treated patients with stereotactic ablative body radiotherapy (SABR) and half treated central tumours with this technique. In fit patients with synchronous solitary brain metastases, 80% would give radical treatment. For curative-intent SCLC, 45–50.4 Gy/25–28# (61%) and 45 Gy/30#/1.5 Gy b.d. (48%) were used. Ninety-four per cent discussed lung cancer patients at multidisciplinary meetings. Contours were peer-reviewed by 74% and 50% for conventional fractionation and SABR respectively. Conclusion: A significant proportion of ROs did not have access to 4D-CT. The majority used 3D image verification and consistently prescribed evidence based doses. A significant number did not participate in peer-review of contours. Practice in IMRT and synchronous oligo-metastatic disease is variable and should be an area of future research. Utilising survey findings, FROLIC is developing consensus recommendations to guide practice. Key words: 4D-CT; IMRT; lung cancer; oligo-metastatic; patterns of practice; SABR.

© 2016 The Royal Australian and New Zealand College of Radiologists

1

Journal of Medical Imaging and Radiation Oncology

RADIATION ONCOLOGY—ORIGINAL ARTICLE

SM Islam et al.

Introduction Radiation therapy plays an important role in the management of all stages of lung cancer for both cure and palliation.1 Despite advances in chemotherapy, surgery and radiation therapy technique, improvements in overall survival have been modest. From 1982–1986 to 2007–2011, 5-year relative survival for lung cancer has only improved from 8% to 14%.2 Lung cancer remains the leading cause of cancer related morbidity and mortality in the Australian population.2 Despite lung cancer’s prevalence and associated mortality rate, prior studies show significant variability in practice,3,4 not only on a national basis, but also within the same state or province. Detailed surveys specific to Australia and New Zealand lung radiation therapy practice were last conducted in 2007.5 Since then, technologies in lung cancer radiation therapy have continued to evolve, and there has been increasing adoption of four-dimensional computed tomography (4D-CT) for tumour motion management and advanced treatment techniques such as stereotactic ablative body radiotherapy (SABR) and intensity-modulated radiation therapy (IMRT). The aim of this study was to evaluate current patterns of care with respect to radiation therapy management of lung cancer throughout Australia and New Zealand to highlight areas for future research and quality improvement.

Methods In October 2014, a web-based survey designed using the Surveymonkeyâ software (SurveyMonkey, Palo Alto, CA, USA) was developed to understand how Radiation Oncologists (ROs) (both subspecialists and generalists) in Australia and New Zealand manage lung cancer. To reach the widest possible audience, ROs in all 62 departments from Australia and New Zealand were emailed utilising the RANZCR database, with replies obtained from all states and territories. These data are to date the most complete representation yet of practice in Australia and New Zealand. There were a total of 36 questions, reviewed by lung cancer subspecialist ROs at the inaugural RANZCR Faculty of Radiation Oncology Lung Interest Cooperative (FROLIC) meeting in September 2014. Questions were designed to obtain information on demographics, simulation techniques, treatment planning, motion management, quality assurance measures, SABR practice, management of oligometastatic disease and small cell lung cancer treatment. The survey was intentionally structured to keep average completion time to within 15 min, thereby reducing “survey fatigue” while maintaining an acceptable completion rate. Question logic software was used to target only ROs who treated lung cancer, and again later in the survey to direct only those who used SABR to an additional set of questions and scenarios relating to SABR.

2

To protect participant privacy, no department information was collected; however, demographic information on State/Territory of practice, rural vs metropolitan and whether the RO was a subspecialist in lung cancer was collected. A subspecialist in lung cancer was defined by FROLIC as any RO with a special interest in treating lung cancer or for whom treating thoracic malignancies constitutes the majority of their practice. Ethics approval was not required for the study. Responses were collected from 24th October 2014 to 17th November 2014. A descriptive analysis was performed with the results presented as a percentage of evaluable responses. If the survey was partially completed, a response was considered evaluable if the question was completed.

Results Sixty-two responses were received, representing 16% of all Australian and New Zealand ROs registered with the Royal Australian and New Zealand College of Radiologists (RANZCR). Of these, 57 (92%) respondents treated patients with lung cancer and were eligible for analysis. Forty-five of 57 respondents (79%) answered all questions in the survey. Sixty-one per cent of respondents were from metropolitan centres and 39% from regional centres. The demographics of respondents by State/Territory are shown in Fig. 1. Sixty per cent of respondents indicated they were subspecialists in lung cancer. Respondents individually saw a mean of 57 and median of 40 (range 5–200) new lung cancer patients per year. Subspecialists saw a mean of 74 (range 15–200) as compared to 30 (range 5–75) for non-subspecialists.

Participation in trials Sixty per cent of respondents reported they were participating in a lung cancer clinical trial (not specific to radiation therapy). Of the 23 respondents not participating in clinical trials, the key reasons cited were lack of radiation specific trials (38%) lack of trial support staff/infrastructure (25%) or not being equipped for SABR (19%). Nineteen per cent of respondents said their research unit was still being set up at the time. Only one respondent (6%) cited lack of funding as a barrier to participation in clinical trials.

Simulation and planning in non-small cell lung cancer Tumour motion management: radical intent ROs were asked to select all options they used in practice for motion management. Fifty-eight per cent of respondents used 4D-CT, while 35% used free breathing CT. Positron emission tomography-computed tomography


Lung radiation therapy patterns of practice: ANZ

Fig. 1. Demographics of Respondents by State/ Territory

30.0%

28.1% 24.6%

Percentage of respondents

25.0%

20.0% 17.5% 15.8%

15.0%

10.0%

5.0%

3.5%

3.5% 1.8%

0.0%

ACT

NSW

NT

3.5% 1.8%

QLD

SA

TAS

VIC

WA

New Zealand

State/Territory

(PET-CT) was commonly used with 23% using PET-CT simulation and 44% PET-CT-fusion. Thirteen per cent used gated CT for planning and 12% used peak inspiration and expiration scans. Fluoroscopy was rarely used (2%). One respondent used PET fusion and performed three simulation CT scans, planning treatment using the CT which showed the best correlation of the GTV to the PET scan.

Tumour motion management: palliative intent For palliative patients, ROs were asked to select all options they would use for motion management. The vast majority of ROs used free breathing CT (81%). About 23% used PET fusion if available and 23% used the same scans as for curative patients.

Respiratory gating techniques Most ROs did not routinely use respiratory gating. Fiftysix per cent of respondents had never utilised breath hold or respiratory gating during simulation. Of the 44% who had used it during simulation, 21% regularly used respiratory gating and 9% breath hold. One respondent (2%) used both techniques for all patients.

Stockholm, Sweden), convolution and superposition/convolution. Less common responses included “Type B algorithms” which take into account lateral electron dose equilibrium (and would include all the algorithms specified above even if the specific algorithm was not specified).

Palliative. In palliative treatments, respondents largely used the same dose calculation algorithms as selected for curative treatments. However, Monte Carlo was employed less often (25%), so that AAA (27%) and other point kernel techniques (29%) predominated.

Treatment Treatment Techniques: Curative Intent ROs were asked what techniques they used to treat patients with curative intent; and allowed to select any that applied. 3D-CRT was used by 98%, with 34% having used IMRT and 18% VMAT. Despite one centre having Cyberknifeâ available, no respondents reported using this for lung treatment. Two respondents (4%) noted treating patients using helical tomotherapy.

Dose calculation algorithms Treatment verification Curative. Monte Carlo (35%) was the most common dose calculation algorithm used in curative treatments, followed by Anisotropic Analytical Algorithm (Varian Medical Systems, Palo Alto, CA, USA) (27%) and other point kernel dose algorithms (25%) i.e. collapsed cone convolution, CMS XIO (Computerized Medical Systems, Saint Louis, MO, USA), Raystation (RaySearch Laboratories,


ROs were asked which treatment verification techniques they used for curative treatment. The majority utilised 3D cone beam CT (86%) and KV imaging (72%). Thirty per cent used MV imaging, 12% used 4D cone beam CT and 6% of respondents used ExacTracâ. In terms of imaging frequency, 52% of respondents used daily CBCT (3D or 4D) during treatment. Thirty-two

3

SM Islam et al.

per cent used a daily CBCT initially on Days 1 to 3–6 then weekly. Only 6% would use daily KV with a weekly CBCT and 4% of respondents would use a 4D-CBCT on day 1 and then daily CBCT. The remaining 6% stated that their practice varies.

Implanted fiducial markers No respondents used fiducials during conventional radiation therapy, while 10% would use fiducials during SABR only.

Lung SABR Forty-four per cent of respondents surveyed were treating patients with lung SABR. Respondents were treating a mean of eight patients (range 0–20) cases per year.

Peripheral tumours not close to chest wall ROs were asked to select all dose/fractionation schedules they used for peripheral tumours not close to the chest wall. The most common schedule was 54 Gy/3#, employed by 80% of respondents. About 30% used 48 Gy/4#. The other fractionation schedule used by one respondent was 54 Gy/4#.

Peripheral tumours close to the chest wall With tumours close to the chest wall, 70% responded that they would use 48 Gy/4#, whilst 20% would use 55 Gy/5# and 10% would not treat lesions close to the chest wall.

Central tumours

ROs were asked how they immobilised patients during SABR and allowed to select all options that applied. Sixty-five per cent used a vacuum bag/vac-fix cushion. Fifty per cent also used a dedicated commercial product. Five per cent reported using abdominal compression.

There was significant heterogeneity in practice for central lesions. Fifty per cent of respondents would not treat central lesions. Thirty per cent would use 60 Gy in 8# and 15% prefer 50 Gy/5#. One respondent used 18 Gy/ 1# to treat a central lesion. Three respondents wrote that they would only treat the central zone in specific scenarios (avoiding very central/hilar tumours, or only treating on trial).

Treatment verification during SABR

Conventional treatment

ROs were asked to select all treatment verification options they employed for SABR. Seventy-five per cent used 3D-CBCT and 50% used 4D-CBCT. A smaller number used KV imaging (25%) or Exactrac (15%). All respondents stated that they performed image verification for every SABR fraction, and 60% performed pre, mid and post fraction verification.

For conventional curative intent treatment in node positive and node negative NSCLC, dose/fractionation schedules that were employed are shown in Figs 2 and 3 respectively. Respondents were allowed to choose more than one schedule. In both cases, “Other” responses used doses between 64 Gy/32# and 65 Gy/35#, while one respondent chose 70 Gy/35# in both scenarios.

Immobilisation during SABR

Fig. 2. Dose/fractionation schedules used by ROs for curative node positive NSCLC

Node positive NSCLC Other

Dose/Fractionation

55Gy/20#

50Gy/20#

13%

15%

17%

66Gy/33#

30%

60Gy/30# 0%

96%

20%

40%

60%

80%

100%


4



Fig. 3. Dose/fractionation schedules used by ROs for curative node negative NSCLC

Node negative NSCLC

Dose/Fractionation

Other

9%

SABR

4%

48Gy/12#

4%

50Gy/20#

30%

55Gy/20#

33%

66Gy/33#

46%

60Gy/30#

83%

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%


performance status who had received definitive treatment of the primary tumour.

Large volume disease precluding radical intent radiation therapy Significant variation in practice was observed in this area. Schedules employed are outlined in Fig. 4. Respondents were allowed to choose more than one schedule. “Other” schedules in use by 22% included 16–17 Gy/2#, 36.5 Gy/15#, 45 Gy/15#, 45 Gy/25#, 54 Gy/30#, 50 Gy/20# and one respondent responded that they would use an “isotoxic regimen”.

Brain metastases Figure 5 shows the percentage of respondents willing to treat with radical intent in the setting of a solitary brain metastasis or up to three brain metastases. Respondents were allowed to choose more than one answer.

Systemic oligo-metastases

Oligo-metastatic disease

For patients with systemic oligo-metastases (e.g. adrenal, bone, liver) 35% stated they would offer radical intent treatment to both chest and extrathoracic disease if there was a single systemic metastasis. This reduced markedly to 6.5% for two metastases and 2% for both

The survey results show large variations in practice when offering radical treatment to oligo-metastatic sites. We assessed willingness to offer radical treatment using a series of scenarios involving a fit patient of good

Fig. 4. Dose/fractionation schedules used by ROs for large volume NSCLC precluding curative radiation therapy

Dose/Fractionation

Large volume disease precluding curative radiation therapy 8Gy/1#

17%

39Gy/13#

17%

40Gy/20#

20%

Other

22%

20Gy/5#

54%

36Gy/12#

72%

30Gy/10# 0%

76%

10%

20%

30%

40%

50%

60%

70%

80%



5

SM Islam et al.

Would you offer curative intent treatment for brain oligometastases with no extrathoracic disease? 0%

10%

20%

30%

40%

50%

60%

Fig. 5. Percentage of ROs willing to treat a fit patient of good performance status with a solitary brain metastasis or up to three brain metastases with radical intent if receiving definitive treatment to the primary

37%

Yes in all cases

13% 43%

Yes if primary disease Stage I or II

22% 9%

No

54% 15% 15%

Yes in selected cases

Solitary Brain Metastasis

Up to Three Brain Metastases

three to five or fewer metastases. Sixty-one per cent stated they do not offer definitive treatment in the setting of systemic oligo-metastases.

Quality assurance practices Discussion of new patients at MDT Ninety-four per cent of respondents discussed new lung cancer patients at a multidisciplinary team meeting. Of these, 63% responded that they would refer all new lung cases to an MDT, 11% would refer curative cases only and 20% would refer only selected patients. Specific criteria used by ROs to “select” patients were not further elucidated by the survey.

Small cell lung cancer Limited stage: curative intent fractionation schedules Dose/fractionation schedules employed in limited stage SCLC are shown in Fig. 6. Respondents were allowed to choose more than one schedule.

Peer review meetings Extensive stage: consolidation chest RT

Seventy-four per cent of respondents stated that they have a peer review meeting to evaluate lung contours for conventional fractionation in both curative and palliative patients. Most commonly (44%), these meetings were to discuss selected patients (either with curative or palliative intent). For SABR cases, 50% of respondents had a peer review meeting to discuss contours for patients with both curative and palliative intent.

Sixty-three per cent of ROs stated they would give consolidation chest RT for patients who achieved a complete response to chemotherapy at metastatic sites, 48% if a partial response was achieved at metastatic sites and 24% would not routinely give consolidation chest RT in ES-SCLC. Respondents were allowed to choose more than one option for this scenario.

Fig. 6. Dose/fractionation schedules used by ROs for limited stage SCLC

Curative intent fractionation schedules for Limited Stage SCLC

Dose/Fractionationan

60Gy/30# o.d.

40Gy/15# o.d.

45Gy/30# b.d.

45-50.4Gy/25-28# o.d.

0%

10%

20%

30%

40%

50%

60%

70%


6



Department protocols Only 37% of departments had both a dose and contouring protocol for lung cancer. Thirty-nine per cent relied on external protocols, e.g. EviQ.6 Eleven per cent did not have a protocol for contouring or dose prescription.

Discussion To our knowledge, this is the first patterns of practice survey to evaluate the use of new technologies, quality assurance activities, clinical trial participation and management of oligo-metastases in the context of lung cancer. A North American survey of ASTRO members was conducted in 2014,7 however, this did not address oligo-metastatic disease or quality control activities, e.g. multi-disciplinary tumour board (MDT) participation, contouring guidelines and peer-review. The survey was sent to every Radiation Oncology department in Australia and New Zealand and drew responses from 16% of all ROs. This is a similar response rate to the ASTRO study (20%).7 However as respondents were not asked to identify their centre, there may be disproportionately more responses from certain centres. If considering responders from Australia only; 33%, 29% and 21% were from Victoria, NSW and QLD respectively. Comparison using Medicare data8 of geographic distribution of the Australian Radiation Oncology Workforce for the same states was 30%, 34% and 18% respectively. This suggests responses were fairly distributed by State/Territory. The survey shows that most practitioners follow evidence-based practice where evidence exists. This may be affected by resource or patient factors for example in SCLC, where many chose 40 Gy/15# once daily as per Murray et al.9 rather than a b.d. regimen and in node negative locally advanced NSCLC where many use hypofractionation. It is likely that hypofractionated regimens are used in those of poorer performance status but the question was not detailed enough to elucidate this with certainty. In keeping with results of RTOG 0617,10 dose escalation beyond 60 Gy for locally advanced NSCLC was uncommon. The use of respiratory gating was uncommon and when used, was usually for selected patients only. This reflects uncertainty around the potential benefits of gating, its resource demands and the complexity of its execution.11 A significant percentage of ROs (42%) did not have access to 4D-CT. In this subgroup, the majority used PET/CT to account for tumour motion. Current Australian lung cancer guidelines recommend either 4D-CT or PET to assess tumour motion.1 The survey also demonstrates significant uptake of newer radiation therapy techniques such as IMRT (34%) and VMAT (18%) among Australia and New Zealand ROs despite the paucity of high-level clinical evidence demonstrating its


benefit over 3D-CRT. The current evidence for IMRT in the setting of lung cancer is limited to those retrospective in nature. However a retrospective review from MD Anderson Cancer Centre showed 4DCRT/IMRT resulted in a significant reduction in toxicity (particularly freedom from Grade ≥3 radiation pneumonitis) and a significant improvement in overall survival (OS) compared with CT/ 3D-CRT.12 As the 3D-CRT arm did not use 4D-CT planning, it is unclear what proportion of this benefit was due to use of the IMRT technique. Conversely, the SEER comparative effectiveness study showed no difference in survival between 3D conformal and IMRT, and no difference in toxicities.13 These current uncertainties in the evidence have lead to a recent article14 highlighting the need for prospective comparative studies. When questioned about current gaps in evidence, the survey demonstrated significant interest in evaluating the clinical benefit of IMRT vs 3D-CRT in a trial setting. SABR for early stage lung cancer is widely practised (44%), and there is general consensus on dose for peripheral tumours far from the chest wall (54 Gy/3# in 80%) and those close to the chest wall (48 Gy/4# in 70%). The variety of dose schedules for central tumours reflects the paucity of high level data and variation in published studies. The results of the CHISEL trial15 are still awaited. The treatment of central tumours is a controversial area. Just half of SABR practitioners were treating central tumours; generally using a risk adapted approach and published regimens.16 The commonest curative regimen for node positive NSCLC was 60 Gy in 30 fractions although 30% also used 66 Gy in 33 fractions. This practice is concordant with results of published studies.17 Only a single respondent had used 70 Gy. For node negative patients with NSCLC, respondents tended to use hypofractionated doses more frequently than for node positive disease. This finding is probably due to smaller and more peripheral planned target volumes (PTV) in node negative disease. Conversely for palliative patients, the wide range of doses employed reflects variables such as performance status and limitations of lung DVHs. A wide range of schedules was used (most commonly 30 Gy/10# by 76% and 36 Gy/12# by 72%). Thirteen per cent of ROs stated they would also treat to higher doses using an isotoxic dose regimen. The survey demonstrates that the use of higher doses to improve local control and survival in better performance status patients is recognised by Australia and New Zealand ROs even when curative intent treatment cannot be given. Most participants nominated more than one potential dose/fractionation regime for the same clinical scenario, reflecting the practice of using “personalised medicine” to tailor the dose based on patient fitness/OAR constraints etc. Alternatively, this could reflect inconsistency in dose prescription or ROs practicing at more than one site with a differing suite of technologies available at each location. This was not further elucidated by the survey.

7

SM Islam et al.

Eighty per cent of ROs were likely to offer radical intent treatment for a fit patient with solitary brain metastasis and no extrathoracic disease. This is consistent with findings from Flannery et al.18 which showed that median overall survival (OS) for a synchronous solitary brain metastases from NSCLC treated with stereotactic radiosurgery (SRS) was 18 months. A randomised trial (RTOG 95-0819) also suggests a functional autonomy benefit to WBRT + SRS, and survival benefit for one metastasis. In the presence of up to three brain metastases, 46% of ROs would consider treating the brain definitively. The heterogeneity in answers most likely reflects the absence of high level evidence to show a survival benefit in treating three brain metastases. A recent meta-analysis20 suggests a survival benefit in patients aged